Have you identified what all the jumpers do?

[UPDATE 4 / DONE] The journey of Tulip Vision Line DC 486sx TC35

There is bound to be 1 or 2 near the clock generator for setting FSB

Tulip Vision Line DC 486dx TC38 Restoration [Update 5] Finished

Re: Dutch members: Tulip Vision Line dt 4/66 (486) CPU jumper help?

Yep, gotta modify the signals to the clock generator, Tulip had 5 different motherboards variations for your model

J16 = unknown, closed 1-2 and 3-4. No difference.
J17 = closed 1-2, no difference. 2-3 must always be closed according to the manual
J18 = when Overdrive R installed, closed since it's a DX4ODPR100 I guess. When open, doesn't boot.
J21 = closed, no difference. X2 looks like an RTC oscillator.

Thanks for replying! 😀

Looking on the retroweb, the board shots aren't great but I think I understand how the clock generator is set up now, the pictures were high enough res that I can make out where the traces are at least.
The ICD2023sc has three pins that are either pulled high or low to select which frequency to run at. Let's ignore the programmable frequency modes that the clock generator can operate in, I don't think it's doing that:

Currently you've got 25MHz FSB correct? So it should be the 50MHz value since FSB is usually clk divided by 2 for actual CPU FSB:

Zooming all the way into this picture on the retroweb I could just about make out that there are joins between
R40 & R41, which seems to link to S2
R42 & R43, looks like that goes to S1
R44 & R45, looks like that goes to S0

My thinking is that a resistor put at R40 will pull the signal low giving S2 a value of "0", while putting a resistor at R41 will pull it high giving it a value of "1"
The layout doesn't make sense for them all to be + then - in order or whatever, maybe the locations are swapped for convenience of PCB layout.

Can you measure the upper side of the resistors & pads present at the top where you see the silkscreen for R41, just by the pins of that ZIPP video memory. Those should link to either 5V or Ground.

If R44 & R45 are linked, potentially you can just move the resistor from R44 onto R45 and that should give you 33MHz FSB. But it's a good idea to verify this first 😀

That is some really valuable information! I'll get to it as soon as possible and of course let you know the outcome.

Update: added clearer picture of the resistors.

Hmm, resistor R44 is already at R45. So I'm going to move it back to R44. 😁

It's an educated guess that the board from your picture was already configured for 33MHz external speed CPU's, i.e. an ordinary DX 33.

Measuring pin 10 (CPUCLK), and yes indeed 50 MHz:

Well, looks like it's working. I'm not 100% happy with the soldering, but those resistors are _tiny_.

That proves again that I am not an SMT pick and place machine.

What matters, is the result!

And we have a full speed DX4 at work!

A very big thanks to Thermalwrong for pointing, no kicking me in the right direction!

That's great 😀 Thanks for the hi-res pic of the PLL area, that did confirm that what I said looks to be correct. Those resistors look to be 33 ohms so they're not far from being jumpers.

Interesting that the picture I was referencing on retroweb already showed the 33MHz PLL configuration, that makes it good for future reference hehe.